A new method for DNA sequencing error verification and correction via an on-disk index tree

Abstract

Existing sequencing error correction techniques demand large expensive memory space. In this work, we introduce a new disk-based sequencing error correction method to solve the problem. The key idea is to utilize a special on-disk index structure, called the BoND-tree, to store and access a large set of k-mers and their associated metadata on disk. With the BoND-tree, a set of special box queries to retrieve the relevant k-mers and their counts are efficiently processed. A comprehensive voting mechanism is adopted to determine and correct an erroneous base in a genome sequence. Experiments demonstrate that the proposed method is quite promising in verifying and correcting sequencing errors in terms of accuracy and scalability.